fix(ecstore): reject stalled ODM pagination before merging pages (#7164)

* fix(odm): reject non-progressing listing cursors

* docs(odm): clarify folded source probe pagination

* test(odm): match SDK bucket-root listing requests

* test(ecstore): match sealed context fixture map type

* test(odm): use app facade for listing wire types

* fix(odm): resolve pagination Clippy failures
This commit is contained in:
Zhengchao An
2026-09-05 15:17:25 +08:00
committed by GitHub
parent a6589c19e3
commit 971f9acdf4
6 changed files with 763 additions and 37 deletions
@@ -86,7 +86,12 @@ impl BreakerVerdict {
Some(SourceError::Throttled | SourceError::Timeout | SourceError::Connect(_) | SourceError::ServerError(_)) => {
BreakerVerdict::Failure
}
Some(SourceError::AccessDenied | SourceError::Unsupported(_) | SourceError::Other(_)) => BreakerVerdict::Neutral,
Some(
SourceError::AccessDenied
| SourceError::Unsupported(_)
| SourceError::InvalidPagination(_)
| SourceError::Other(_),
) => BreakerVerdict::Neutral,
}
}
}
@@ -189,8 +189,8 @@ pub enum SourceListPlan {
/// delimiter — the source's own roll-up boundary matches the request's.
Page { prefix: String },
/// `filter.prefix` reaches past a delimiter, so every key the source could
/// contribute rolls into this one common prefix. One bounded probe listing
/// decides whether it exists; there is nothing to paginate.
/// contribute rolls into this one common prefix. Bounded probes follow
/// empty progressing pages until a key proves existence or the source ends.
Folded { probe_prefix: String, common_prefix: String },
}
@@ -279,6 +279,29 @@ pub struct FetchRequest {
pub token: Option<String>,
}
/// Invalid pagination metadata. Opaque cursor values are never included in errors.
#[derive(Clone, Copy, Debug, PartialEq, Eq, thiserror::Error)]
pub enum ListPageError {
#[error("truncated listing has no continuation token")]
Missing,
#[error("truncated listing has an empty continuation token")]
Empty,
#[error("truncated listing repeats a continuation token")]
Repeated,
}
pub(crate) fn validate_list_page(is_truncated: bool, token: Option<&str>, next_token: Option<&str>) -> Result<(), ListPageError> {
if is_truncated {
match next_token {
None => return Err(ListPageError::Missing),
Some("") => return Err(ListPageError::Empty),
Some(next) if Some(next) == token => return Err(ListPageError::Repeated),
Some(_) => {}
}
}
Ok(())
}
#[derive(Debug, Default)]
struct SideState {
start: SideCursor,
@@ -364,6 +387,11 @@ impl ListThroughMerger {
/// or `filter.prefix` excludes it.
pub fn disable_source(&mut self) {
self.source.disabled = true;
// A refill can fail after a valid first page. A local-only response
// must discard both that source payload and its ordering horizon.
self.source.entries.clear();
self.source.pages.clear();
self.source.more = false;
}
pub fn next_fetch(&self) -> Option<FetchRequest> {
@@ -378,7 +406,13 @@ impl ListThroughMerger {
/// Records one fetched page. `entries` must be sorted by `name` and already
/// filtered with [`Self::accepts`]; the caller keeps the matching payloads
/// in the same order.
pub fn push_page(&mut self, side: MergeSide, entries: Vec<ListEntryKey>, is_truncated: bool, next_token: Option<String>) {
pub fn push_page(
&mut self,
side: MergeSide,
entries: Vec<ListEntryKey>,
is_truncated: bool,
next_token: Option<String>,
) -> Result<(), ListPageError> {
let state = match side {
MergeSide::Local => &mut self.local,
MergeSide::Source => &mut self.source,
@@ -387,15 +421,19 @@ impl ListThroughMerger {
Some(last) => last.next_token.clone(),
None => state.start.token.clone(),
};
// A truncated page without a cursor cannot be continued; treating the
// side as finished is the only alternative to looping on it forever.
state.more = is_truncated && next_token.is_some();
validate_list_page(is_truncated, token.as_deref(), next_token.as_deref())?;
// Also reject a cycle through an earlier page in this bounded fetch.
if is_truncated && state.pages.iter().any(|page| page.token == next_token) {
return Err(ListPageError::Repeated);
}
state.more = is_truncated;
state.pages.push(FetchedPage {
token,
count: entries.len(),
next_token: is_truncated.then_some(next_token).flatten(),
});
state.entries.extend(entries);
Ok(())
}
pub fn finish(self) -> MergeOutcome {
@@ -599,9 +637,15 @@ mod tests {
let (entries, truncated, next) = reference_page(keys, prefix, delimiter, fetch.token.as_deref(), max_keys);
let kept: Vec<ListEntryKey> = entries.into_iter().filter(|entry| merger.accepts(&entry.name)).collect();
buffers[usize::from(fetch.side == MergeSide::Source)].extend(kept.iter().cloned());
merger.push_page(fetch.side, kept, truncated, next);
merger
.push_page(fetch.side, kept, truncated, next)
.expect("reference provider pages must advance");
}
let outcome = merger.finish();
assert_eq!(outcome.is_truncated, outcome.next_token.is_some());
if outcome.is_truncated {
assert_ne!(outcome.next_token, token, "every truncated merged page must make progress");
}
page_sizes.push(outcome.picks.len());
for pick in &outcome.picks {
let entry = buffers[usize::from(pick.side == MergeSide::Source)][pick.index].clone();
@@ -616,11 +660,25 @@ mod tests {
}
fn expected(local: &[String], source: &[String], prefix: &str, delimiter: Option<&str>) -> Vec<ListEntryKey> {
let mut all: Vec<String> = local.iter().chain(source.iter()).cloned().collect();
all.sort();
all.dedup();
let (entries, _, _) = reference_page(&all, prefix, delimiter, None, usize::MAX);
entries
// This oracle builds the complete namespace independently of the
// provider's page/marker helper and the production merger.
let mut namespace = std::collections::BTreeMap::new();
for key in local.iter().chain(source) {
let Some(suffix) = key.strip_prefix(prefix) else {
continue;
};
if let Some(delimiter) = delimiter.filter(|delimiter| !delimiter.is_empty())
&& let Some((directory, _)) = suffix.split_once(delimiter)
{
namespace.insert(format!("{prefix}{directory}{delimiter}"), true);
continue;
}
namespace.insert(key.clone(), false);
}
namespace
.into_iter()
.map(|(name, is_prefix)| ListEntryKey { name, is_prefix })
.collect()
}
#[test]
@@ -662,7 +720,9 @@ mod tests {
token: None
})
);
merger.push_page(MergeSide::Local, vec![ListEntryKey::object("a")], false, None);
merger
.push_page(MergeSide::Local, vec![ListEntryKey::object("a")], false, None)
.expect("local EOF is valid");
assert_eq!(merger.next_fetch(), None);
let outcome = merger.finish();
assert_eq!(outcome.picks.len(), 1);
@@ -683,12 +743,14 @@ mod tests {
};
let mut merger = ListThroughMerger::new(1, Some(&resume));
merger.disable_source();
merger.push_page(
MergeSide::Local,
vec![ListEntryKey::object("b"), ListEntryKey::object("c")],
true,
Some("local-2".to_string()),
);
merger
.push_page(
MergeSide::Local,
vec![ListEntryKey::object("b"), ListEntryKey::object("c")],
true,
Some("local-2".to_string()),
)
.expect("local cursor advances");
let outcome = merger.finish();
assert!(outcome.is_truncated);
let token = outcome.next_token.expect("truncated page carries a token");
@@ -698,6 +760,212 @@ mod tests {
assert_eq!(token.local.as_deref(), Some("local-1"), "a partly read page is re-listed");
}
#[test]
fn truncated_pages_require_a_nonempty_advancing_cursor() {
for side in [MergeSide::Local, MergeSide::Source] {
for entries in [vec![], vec![ListEntryKey::object("a")]] {
for (next, expected) in [
(None, Err(ListPageError::Missing)),
(Some(""), Err(ListPageError::Empty)),
(Some("stuck"), Err(ListPageError::Repeated)),
(Some("advances"), Ok(())),
] {
let resume = ListThroughToken::new(
SideCursor {
token: Some("stuck".into()),
done: false,
},
SideCursor {
token: Some("stuck".into()),
done: false,
},
None,
);
let mut merger = ListThroughMerger::new(2, Some(&resume));
let result = merger.push_page(side, entries.clone(), true, next.map(str::to_string));
assert_eq!(result, expected, "{side:?}, {entries:?}, {next:?}");
let state = if side == MergeSide::Local {
&merger.local
} else {
&merger.source
};
assert_eq!(state.pages.len(), usize::from(result.is_ok()), "invalid page must not be accepted");
}
}
}
}
#[test]
fn repeated_empty_cursor_is_rejected_before_an_identical_page_can_escape() {
let resume = ListThroughToken::new(
SideCursor { token: None, done: true },
SideCursor {
token: Some("stuck".into()),
done: false,
},
None,
);
let mut merger = ListThroughMerger::new(2, Some(&resume));
assert_eq!(
merger.next_fetch(),
Some(FetchRequest {
side: MergeSide::Source,
token: Some("stuck".into())
})
);
assert_eq!(
merger.push_page(MergeSide::Source, vec![], true, Some("stuck".into())),
Err(ListPageError::Repeated)
);
}
#[test]
fn empty_pages_may_advance_within_the_fetch_budget_until_eof() {
let mut merger = ListThroughMerger::new(2, None);
merger.push_page(MergeSide::Local, vec![], false, None).expect("local EOF");
for next in ["opaque-z", "opaque-a"] {
assert_eq!(merger.next_fetch().expect("bounded source fetch").side, MergeSide::Source);
merger
.push_page(MergeSide::Source, vec![], true, Some(next.into()))
.expect("opaque cursor advances regardless of sort order");
}
assert!(merger.next_fetch().is_none(), "two source fetches exhaust the request budget");
let outcome = merger.finish();
assert!(outcome.picks.is_empty());
assert!(outcome.is_truncated);
let token = outcome.next_token.expect("empty progressing page has a cursor");
assert_eq!(token.source.as_deref(), Some("opaque-a"));
let mut merger = ListThroughMerger::new(2, Some(&token));
assert_eq!(merger.next_fetch().expect("source resumes").token.as_deref(), Some("opaque-a"));
merger
.push_page(MergeSide::Source, vec![ListEntryKey::object("result")], false, None)
.expect("source EOF");
let outcome = merger.finish();
assert_eq!(
outcome.picks,
vec![MergePick {
side: MergeSide::Source,
index: 0
}]
);
assert!(!outcome.is_truncated);
assert!(outcome.next_token.is_none());
}
#[test]
fn a_cursor_cycle_inside_the_fetch_budget_is_rejected() {
let resume = ListThroughToken::new(
SideCursor { token: None, done: true },
SideCursor {
token: Some("first".into()),
done: false,
},
None,
);
let mut merger = ListThroughMerger::new(2, Some(&resume));
merger
.push_page(MergeSide::Source, vec![], true, Some("second".into()))
.expect("first page advances");
assert_eq!(
merger.push_page(MergeSide::Source, vec![], true, Some("first".into())),
Err(ListPageError::Repeated)
);
}
#[test]
fn source_refill_failure_discards_buffered_source_entries_and_horizon() {
let mut merger = ListThroughMerger::new(2, None);
merger
.push_page(MergeSide::Local, vec![ListEntryKey::object("z")], false, None)
.expect("local EOF");
merger
.push_page(MergeSide::Source, vec![ListEntryKey::object("a")], true, Some("stuck".into()))
.expect("first source page advances");
assert_eq!(merger.next_fetch().expect("source refill is required").token.as_deref(), Some("stuck"));
assert_eq!(
merger.push_page(MergeSide::Source, vec![], true, Some("stuck".into())),
Err(ListPageError::Repeated)
);
merger.disable_source();
let outcome = merger.finish();
assert_eq!(
outcome.picks,
vec![MergePick {
side: MergeSide::Local,
index: 0
}]
);
assert!(!outcome.is_truncated);
assert!(outcome.next_token.is_none());
}
#[test]
fn list_through_static_namespace_boundary_matrix() {
let corpus = [
"a",
"a/",
"a/b",
"a/b/child",
"a0",
"b",
"b/leaf",
"quote\"&<",
"space key",
"z",
"é",
"中/文",
];
for count in [0, 1, 3, 4, corpus.len()] {
let keys: Vec<String> = corpus[..count].iter().map(|key| (*key).to_string()).collect();
for placement in 0..3 {
let (local, source): (Vec<_>, Vec<_>) =
keys.iter()
.enumerate()
.fold((vec![], vec![]), |(mut local, mut source), (index, key)| {
if placement != 1 || index % 2 == 0 {
local.push(key.clone());
}
if placement != 0 || index % 2 == 0 {
source.push(key.clone());
}
(local, source)
});
for prefix in ["", "a", "a/", "中/"] {
for delimiter in [None, Some("/")] {
for max_keys in [1, 3, 4] {
let oracle = expected(&local, &source, prefix, delimiter);
let (emitted, sizes) = walk(&local, &source, prefix, delimiter, max_keys);
assert_eq!(
emitted.iter().map(|(entry, _)| entry.clone()).collect::<Vec<_>>(),
oracle,
"count={count}, placement={placement}, prefix={prefix}, delimiter={delimiter:?}, max={max_keys}"
);
let expected_sizes: Vec<_> = if oracle.is_empty() {
vec![0]
} else {
oracle.chunks(max_keys).map(<[ListEntryKey]>::len).collect()
};
assert_eq!(sizes, expected_sizes, "exact max and max+1 boundaries must agree");
}
}
}
}
}
}
#[test]
fn list_through_large_overlap_walk_keeps_all_5300_keys() {
let source: Vec<_> = (0..5000).map(|index| format!("k{index:05}")).collect();
let local: Vec<_> = (4800..5300).map(|index| format!("k{index:05}")).collect();
let (emitted, sizes) = walk(&local, &source, "", None, 333);
assert_eq!(emitted.len(), 5300);
for (index, (entry, side)) in emitted.iter().enumerate() {
assert_eq!(entry.name, format!("k{index:05}"));
assert_eq!(*side, if index >= 4800 { MergeSide::Local } else { MergeSide::Source });
}
assert_eq!(sizes, [vec![333; 15], vec![305]].concat());
}
#[test]
fn token_round_trips_and_rejects_tampering() {
let token = ListThroughToken::new(
@@ -796,7 +1064,10 @@ mod tests {
}
proptest! {
#![proptest_config(ProptestConfig::with_cases(256))]
#![proptest_config(ProptestConfig {
rng_seed: proptest::test_runner::RngSeed::Fixed(0xec5706),
..ProptestConfig::with_cases(256)
})]
/// Full pagination of a merged listing equals the sorted, deduplicated
/// union of both sides, with every shared key served by local, and no
@@ -25,6 +25,7 @@
//! Client-supplied `If-*`, `Authorization`, `Host` and SSE-C headers are never
//! forwarded: v1 rejects SSE-C source objects outright.
use super::list_through::{ListPageError, validate_list_page};
use crate::bucket::remote_s3_client::{
PathStyle, RemoteCredentials, RemoteS3ClientError, RemoteS3EndpointSpec, RemoteS3RetryPolicy, build_remote_s3_config,
};
@@ -223,6 +224,8 @@ pub enum SourceError {
ServerError(u16),
#[error("unsupported source object: {0}")]
Unsupported(String),
#[error("invalid source listing: {0}")]
InvalidPagination(#[from] ListPageError),
#[error("source request failed: {0}")]
Other(String),
}
@@ -245,6 +248,7 @@ impl SourceError {
SourceError::Connect(_) => "connect",
SourceError::ServerError(_) => "server_error",
SourceError::Unsupported(_) => "unsupported",
SourceError::InvalidPagination(_) => "invalid_pagination",
SourceError::Other(_) => "other",
}
}
@@ -714,6 +718,7 @@ impl SourceClient {
..*request
})
.await?;
validate_list_page(page.is_truncated, request.continuation_token, page.next_continuation_token.as_deref())?;
page.objects = page
.objects
.into_iter()
@@ -800,11 +805,6 @@ impl SourceBackend for S3SourceBackend {
let is_truncated = output.is_truncated.unwrap_or(false);
let next_continuation_token = output.next_continuation_token;
if is_truncated && next_continuation_token.is_none() {
return Err(SourceError::Other(
"source reported a truncated listing without a continuation token".to_string(),
));
}
let objects = output
.contents
.unwrap_or_default()
@@ -1274,7 +1274,9 @@ mod tests {
<CommonPrefixes><Prefix>data/photos/</Prefix></CommonPrefixes>
<CommonPrefixes><Prefix>outside/</Prefix></CommonPrefixes>
</ListBucketResult>"#;
let (client, requests) = scripted_client(&spec(Some("data/")), vec![ok(Vec::new(), body), ok(Vec::new(), body)]).await;
let next_body = body.replace("data/opaque", "data/next");
let (client, requests) =
scripted_client(&spec(Some("data/")), vec![ok(Vec::new(), body), ok(Vec::new(), &next_body)]).await;
let first = client
.list_page(&SourceListRequest {
prefix: Some("photos/"),
@@ -1336,7 +1338,104 @@ mod tests {
.list_objects_v2(None, None, 10)
.await
.expect_err("truncated page without token is corrupt");
assert!(matches!(err, SourceError::Other(_)), "{err:?}");
assert!(matches!(err, SourceError::InvalidPagination(ListPageError::Missing)), "{err:?}");
}
#[tokio::test]
async fn list_page_validates_s3_cursor_progress_before_mapping_entries() {
for contents in ["", "<Contents><Key>data/a</Key><Size>1</Size></Contents>"] {
for (truncated, next, expected) in [
(true, None, Some(ListPageError::Missing)),
(true, Some(""), Some(ListPageError::Empty)),
(true, Some("stuck"), Some(ListPageError::Repeated)),
(true, Some("opaque-next"), None),
(false, None, None),
(false, Some("stuck"), None),
] {
let next_xml = next
.map(|next| format!("<NextContinuationToken>{next}</NextContinuationToken>"))
.unwrap_or_default();
let body = format!(
"<ListBucketResult xmlns=\"http://s3.amazonaws.com/doc/2006-03-01/\"><IsTruncated>{truncated}</IsTruncated>{next_xml}{contents}</ListBucketResult>"
);
let (client, requests) = scripted_client(&spec(Some("data/")), vec![ok(Vec::new(), &body)]).await;
let result = client
.list_page(&SourceListRequest {
continuation_token: Some("stuck"),
max_keys: 2,
..Default::default()
})
.await;
match expected {
Some(expected) => {
let error = result.expect_err("malformed pagination must fail at the provider boundary");
assert!(
matches!(&error, SourceError::InvalidPagination(actual) if *actual == expected),
"{error:?}"
);
assert_eq!(error.class_label(), "invalid_pagination");
assert!(!error.is_retryable());
assert!(!error.to_string().contains("stuck"), "errors must not echo opaque tokens");
}
None => {
let page = result.expect("progressing empty/nonempty pages and EOF are valid");
assert_eq!(page.is_truncated, truncated);
assert_eq!(page.next_continuation_token.as_deref(), next);
assert_eq!(page.objects.len(), usize::from(!contents.is_empty()));
if let Some(object) = page.objects.first() {
assert_eq!(object.key, "a");
}
}
}
let requests = recorded(&requests);
assert_eq!(requests.len(), 1, "invalid pagination must not be retried");
assert!(requests[0].uri.contains("continuation-token=stuck"));
}
}
}
struct ListOnlyBackend(SourcePage);
#[async_trait::async_trait]
impl SourceBackend for ListOnlyBackend {
async fn list(&self, request: &SourceListRequest<'_>) -> Result<SourcePage, SourceError> {
assert_eq!(request.continuation_token, Some("stuck"), "opaque cursors reach every provider unchanged");
Ok(self.0.clone())
}
async fn head(&self, _key: &str) -> Result<SourceHead, SourceError> {
panic!("unexpected HEAD in list test")
}
async fn get(&self, _key: &str, _range: Option<&HTTPRangeSpec>) -> Result<SourceGet, SourceError> {
panic!("unexpected GET in list test")
}
async fn tagging(&self, _key: &str) -> Result<HashMap<String, String>, SourceError> {
panic!("unexpected tagging in list test")
}
async fn probe(&self) -> Result<(), SourceError> {
panic!("unexpected probe in list test")
}
}
#[tokio::test]
async fn list_page_validates_non_s3_provider_cursors_at_the_common_boundary() {
for (next, expected) in [
(None, ListPageError::Missing),
(Some(""), ListPageError::Empty),
(Some("stuck"), ListPageError::Repeated),
] {
let mut client = prefix_client(Some("data/".into()));
client.backend = Box::new(ListOnlyBackend(SourcePage {
is_truncated: true,
next_continuation_token: next.map(str::to_string),
..Default::default()
}));
let error = client
.list_objects_v2(None, Some("stuck"), 2)
.await
.expect_err("all providers must advance pagination");
assert!(matches!(error, SourceError::InvalidPagination(actual) if actual == expected));
}
}
const TAGGING_BODY: &str = r#"<?xml version="1.0" encoding="UTF-8"?>
@@ -177,7 +177,7 @@ impl From<&SourceError> for PullFailureReason {
SourceError::Connect(_) => PullFailureReason::SourceConnect,
SourceError::ServerError(_) => PullFailureReason::SourceServerError,
SourceError::Unsupported(_) => PullFailureReason::SourceUnsupported,
SourceError::Other(_) => PullFailureReason::SourceOther,
SourceError::InvalidPagination(_) | SourceError::Other(_) => PullFailureReason::SourceOther,
}
}
}
+354 -5
View File
@@ -252,8 +252,16 @@ pub(crate) async fn merged_list_objects_v2(
.filter(|entry| merger.accepts(&entry.key().name))
.collect();
let keys: Vec<ListEntryKey> = kept.iter().map(SideEntry::key).collect();
if let Err(error) = merger.push_page(fetch.side, keys, is_truncated, next_token) {
match fetch.side {
MergeSide::Source => {
degrade_or_fail(&mut merger, &mut degraded, policy.source_error, "invalid_pagination")?;
continue;
}
MergeSide::Local => return Err(S3Error::with_message(S3ErrorCode::InternalError, error.to_string())),
}
}
buffers[usize::from(fetch.side == MergeSide::Source)].extend(kept.into_iter().map(Some));
merger.push_page(fetch.side, keys, is_truncated, next_token);
}
let outcome = merger.finish();
@@ -340,10 +348,12 @@ async fn fetch_source_page(
continuation_token: token,
max_keys: params.max_keys,
},
// Everything under `filter.prefix` rolls into one common prefix, so a
// single bounded listing settles whether it exists.
// Everything under `filter.prefix` rolls into one common prefix. An
// empty truncated probe must still follow its cursor before declaring
// that prefix absent.
SourceListPlan::Folded { probe_prefix, .. } => SourceListRequest {
prefix: Some(probe_prefix.as_str()),
continuation_token: token,
max_keys: 1,
..Default::default()
},
@@ -368,8 +378,8 @@ async fn fetch_source_page(
} else {
Vec::new()
},
false,
None,
!exists && page.is_truncated,
if exists { None } else { page.next_continuation_token },
))
}
_ => {
@@ -417,6 +427,17 @@ async fn local_delete_markers(store: &Arc<ECStore>, bucket: &str, keys: &[String
#[cfg(test)]
mod tests {
use super::*;
use crate::app::bucket_usecase::DefaultBucketUsecase;
use crate::app::gating_test_env::{run_large_stack_test, shared_gating_ecstore};
use crate::app::storage_api::bucket_usecase::bucket::on_demand_migration::{
FilterConfig, OnDemandMigrationConfig, PathStyle, PolicyConfig, Provider, SourceConfig, SourceCredentials, TlsConfig,
};
use crate::app::storage_api::bucket_usecase::s3::{ListObjectsV2Input, ListObjectsV2Output, S3Request, S3Response};
use crate::app::storage_api::test::StoragePutObjReader;
use crate::app::storage_api::test::contract::bucket::{BucketOperations as _, MakeBucketOptions};
use crate::app::storage_api::test::contract::object::ObjectIO as _;
use std::time::Duration;
use tokio::io::{AsyncReadExt, AsyncWriteExt};
fn token(local: Option<&str>, local_done: bool) -> ListThroughToken {
ListThroughToken {
@@ -526,4 +547,332 @@ mod tests {
assert!(degraded);
assert_eq!(merger.next_fetch().map(|fetch| fetch.side), Some(MergeSide::Local));
}
/// Serves exactly the scripted S3 pages and joins every connection before
/// returning. A source retry or unexpected operation fails the test.
async fn scripted_list_source(pages: Vec<String>) -> (String, tokio_util::task::AbortOnDropHandle<Vec<String>>) {
let listener = tokio::net::TcpListener::bind("127.0.0.1:0")
.await
.expect("bind listing source");
let address = listener.local_addr().expect("listing source address");
let server = tokio::spawn(async move {
let mut requests = Vec::new();
for body in pages {
let (mut stream, _) = listener.accept().await.expect("accept source listing");
let mut request = Vec::new();
let mut chunk = [0; 4096];
while !request.windows(4).any(|window| window == b"\r\n\r\n") {
let count = stream.read(&mut chunk).await.expect("read signed listing request");
assert!(count > 0, "source request must include complete headers");
request.extend_from_slice(&chunk[..count]);
assert!(request.len() <= 32 * 1024, "listing request headers must be bounded");
}
let first_line = String::from_utf8_lossy(&request)
.lines()
.next()
.expect("request line")
.to_string();
// The SDK joins the bucket endpoint with the LIST operation's `/` path.
assert!(
first_line.starts_with("GET /source-bucket/?"),
"expected a path-style bucket-root LIST request, got {first_line:?}"
);
assert!(first_line.contains("list-type=2"), "expected a ListObjectsV2 query, got {first_line:?}");
requests.push(first_line);
let response = format!(
"HTTP/1.1 200 OK\r\ncontent-type: application/xml\r\ncontent-length: {}\r\nconnection: close\r\n\r\n{body}",
body.len()
);
stream.write_all(response.as_bytes()).await.expect("write source page");
stream.shutdown().await.expect("finish source response");
}
requests
});
(format!("http://{address}"), tokio_util::task::AbortOnDropHandle::new(server))
}
fn source_xml(next: Option<&str>, truncated: bool, key: Option<&str>) -> String {
let next = next
.map(|token| format!("<NextContinuationToken>{token}</NextContinuationToken>"))
.unwrap_or_default();
let contents = key
.map(|key| format!("<Contents><Key>{key}</Key><Size>1</Size></Contents>"))
.unwrap_or_default();
format!(
"<ListBucketResult xmlns=\"http://s3.amazonaws.com/doc/2006-03-01/\"><IsTruncated>{truncated}</IsTruncated>{next}{contents}</ListBucketResult>"
)
}
struct ListThroughTestState {
bucket: String,
module_enabled: bool,
}
impl Drop for ListThroughTestState {
fn drop(&mut self) {
let sys = OnDemandMigrationSys::get();
sys.remove(&self.bucket);
sys.set_module_enabled(self.module_enabled);
}
}
async fn source_policy_request(
pages: Vec<String>,
policy: SourceErrorPolicy,
resume_source: Option<&str>,
filter_prefix: Option<&str>,
) -> (S3Result<S3Response<ListObjectsV2Output>>, Vec<String>) {
let store = shared_gating_ecstore().await;
crate::app::runtime_sources::install_test_app_context(Arc::clone(&store)).await;
let bucket = format!("odm-list-{}", uuid::Uuid::new_v4().simple());
store
.make_bucket(&bucket, &MakeBucketOptions::default())
.await
.expect("create list-through bucket");
store
.put_object(
&bucket,
"z-local",
&mut StoragePutObjReader::from_vec(vec![1]),
&StorageObjectOptions::default(),
)
.await
.expect("seed real local listing");
let (endpoint, server) = scripted_list_source(pages).await;
let sys = OnDemandMigrationSys::get();
let _state_guard = ListThroughTestState {
bucket: bucket.clone(),
module_enabled: sys.is_module_enabled(),
};
sys.set_module_enabled(true);
let config = OnDemandMigrationConfig {
version: 1,
enabled: true,
source: SourceConfig {
provider: Provider::Minio,
endpoint: Some(endpoint),
region: "us-east-1".into(),
bucket: "source-bucket".into(),
path_style: PathStyle::Path,
credentials: Some(SourceCredentials {
access_key: "test-access".into(),
secret_key: "test-secret".into(),
session_token: None,
}),
tls: TlsConfig::default(),
},
filter: FilterConfig {
prefix: filter_prefix.map(str::to_string),
..Default::default()
},
policy: PolicyConfig {
list_through: true,
source_error: policy,
..Default::default()
},
};
sys.apply(&bucket, Some(&config)).await;
assert!(
sys.state(&bucket).expect("ODM state installed").client().is_ok(),
"fake source client must build"
);
let continuation_token = resume_source.map(|source| {
let token = ListThroughToken {
t: "odm-list".into(),
v: 1,
local: None,
local_done: false,
source: Some(source.into()),
source_done: false,
last_key: None,
};
base64_simd::STANDARD.encode_to_string(token.encode().as_bytes())
});
let input = ListObjectsV2Input {
bucket,
max_keys: Some(2),
continuation_token,
delimiter: filter_prefix.map(|_| "/".to_string()),
encoding_type: None,
expected_bucket_owner: None,
fetch_owner: None,
optional_object_attributes: None,
prefix: None,
request_payer: None,
start_after: None,
};
let request = S3Request {
input,
method: http::Method::GET,
uri: http::Uri::from_static("/?list-type=2"),
headers: HeaderMap::new(),
extensions: http::Extensions::new(),
credentials: None,
region: None,
service: None,
trailing_headers: None,
};
let result = tokio::time::timeout(
Duration::from_secs(10),
DefaultBucketUsecase::from_global().execute_list_objects_v2(request),
)
.await
.expect("listing must complete within its bounded source budget");
let requests = tokio::time::timeout(Duration::from_secs(5), server)
.await
.expect("source connections must finish")
.expect("source server must not panic");
(result, requests)
}
#[test]
#[serial_test::serial]
fn list_through_invalid_source_pagination_obeys_policy_on_the_handler_path() {
run_large_stack_test("list-through-source-policy", || async {
temp_env::async_with_vars(
[
("RUSTFS_REPLICATION_ALLOW_LOOPBACK_TARGET", Some("true")),
("HTTP_PROXY", None),
("HTTPS_PROXY", None),
("ALL_PROXY", None),
("http_proxy", None),
("https_proxy", None),
("all_proxy", None),
("NO_PROXY", Some("*")),
("no_proxy", Some("*")),
],
async {
for policy in [SourceErrorPolicy::Propagate, SourceErrorPolicy::NotFound] {
for next in [None, Some(""), Some("stuck")] {
for key in [None, Some("a-source")] {
let (result, requests) =
source_policy_request(vec![source_xml(next, true, key)], policy, Some("stuck"), None).await;
assert_eq!(requests.len(), 1, "a malformed source page must not be retried");
assert!(requests[0].contains("continuation-token=stuck"));
assert_source_policy_result(result, policy);
}
}
let (result, requests) = source_policy_request(
vec![
source_xml(Some("stuck"), true, Some("a-source")),
source_xml(Some("stuck"), true, None),
],
policy,
None,
None,
)
.await;
assert_eq!(requests.len(), 2, "the failure must occur during a real refill");
assert!(!requests[0].contains("continuation-token="));
assert!(requests[1].contains("continuation-token=stuck"));
assert_source_policy_result(result, policy);
}
},
)
.await;
});
}
#[test]
#[serial_test::serial]
fn list_through_empty_advancing_source_pages_reach_eof_on_the_handler_path() {
run_large_stack_test("list-through-empty-source-pages", || async {
temp_env::async_with_vars(
[
("RUSTFS_REPLICATION_ALLOW_LOOPBACK_TARGET", Some("true")),
("HTTP_PROXY", None),
("HTTPS_PROXY", None),
("ALL_PROXY", None),
("http_proxy", None),
("https_proxy", None),
("all_proxy", None),
("NO_PROXY", Some("*")),
("no_proxy", Some("*")),
],
async {
for filter_prefix in [None, Some("photos/2024/")] {
let source_key = if filter_prefix.is_some() {
"photos/2024/a-source"
} else {
"a-source"
};
let (result, requests) = source_policy_request(
vec![
source_xml(Some("opaque-next"), true, None),
source_xml(None, false, Some(source_key)),
],
SourceErrorPolicy::Propagate,
None,
filter_prefix,
)
.await;
assert_eq!(requests.len(), 2, "an empty truncated source page must reach its successor");
assert!(requests[1].contains("continuation-token=opaque-next"));
let response = result.expect("empty progressing source page is valid");
assert!(!response.headers.contains_key("x-rustfs-on-demand-migration-list"));
let output = response.output;
let objects: Vec<_> = output
.contents
.unwrap_or_default()
.into_iter()
.map(|object| object.key.expect("listed object key"))
.collect();
if filter_prefix.is_some() {
assert_eq!(objects, vec!["z-local"]);
assert_eq!(
output
.common_prefixes
.unwrap_or_default()
.into_iter()
.map(|prefix| prefix.prefix.expect("rolled-up prefix"))
.collect::<Vec<_>>(),
vec!["photos/"]
);
} else {
assert_eq!(objects, vec!["a-source", "z-local"]);
assert!(output.common_prefixes.unwrap_or_default().is_empty());
}
assert_eq!(output.key_count, Some(2));
assert_eq!(output.is_truncated, Some(false));
assert!(output.next_continuation_token.is_none());
}
},
)
.await;
});
}
fn assert_source_policy_result(result: S3Result<S3Response<ListObjectsV2Output>>, policy: SourceErrorPolicy) {
match policy {
SourceErrorPolicy::Propagate => {
let error = result.expect_err("propagate must expose malformed pagination");
assert_eq!(error.status_code(), Some(http::StatusCode::FAILED_DEPENDENCY));
assert_eq!(error.code(), &S3ErrorCode::Custom("SourceUnavailable".into()));
assert_eq!(error.message(), Some("invalid_pagination"));
}
SourceErrorPolicy::NotFound => {
let response = result.expect("not_found must preserve the local listing");
assert_eq!(
response
.headers
.get("x-rustfs-on-demand-migration-list")
.expect("local_only header"),
"local_only"
);
let output = response.output;
assert_eq!(
output
.contents
.unwrap_or_default()
.into_iter()
.map(|object| object.key.expect("local key"))
.collect::<Vec<_>>(),
vec!["z-local"]
);
assert_eq!(output.is_truncated, Some(false));
assert_eq!(output.key_count, Some(1));
assert!(output.next_continuation_token.is_none());
}
}
}
}
+5 -3
View File
@@ -29,11 +29,13 @@ pub(crate) fn EndpointServerPools(
pub(crate) mod s3 {
#[cfg(test)]
pub(crate) use s3s::dto::{
BucketVersioningStatus, DeleteMarkerReplication, DeleteMarkerReplicationStatus, Destination, ReplicationConfiguration,
ReplicationRule, ReplicationRuleFilter, ReplicationRuleStatus, ServerSideEncryptionByDefault,
ServerSideEncryptionConfiguration, ServerSideEncryptionRule, Tag, VersioningConfiguration,
BucketVersioningStatus, DeleteMarkerReplication, DeleteMarkerReplicationStatus, Destination, ListObjectsV2Input,
ListObjectsV2Output, ReplicationConfiguration, ReplicationRule, ReplicationRuleFilter, ReplicationRuleStatus,
ServerSideEncryptionByDefault, ServerSideEncryptionConfiguration, ServerSideEncryptionRule, Tag, VersioningConfiguration,
};
pub(crate) use s3s::{S3Error, S3ErrorCode, S3Result};
#[cfg(test)]
pub(crate) use s3s::{S3Request, S3Response};
}
pub(crate) mod admin {